Authors
K.D. Edwards, R.M. Wagner, J.B. Green Jr., C.S. Daw
Publication
Proceedings: The Third Joint Meeting of the U.S. Sections of the
Combustion Institute, Chicago, IL, March 2003
Abstract
A
predictive feedback control strategy based on low-order mapping of
system behavior is applied to reduce cyclic dispersion in a lean
spark-ignition combustion model during transient operation. History of
the low-dimensional, steady-state behavior of the model is used to
construct a series of maps that predict future combustion events. We
explore the effectiveness of a control strategy which uses these static
maps to predict future undesirable combustion events and apply
appropriate control perturbations to steer the system toward more
desirable behavior. Successful control is demonstrated as a transient
fueling is applied to the model which, if uncontrolled, would lead to
the development of combustion instabilities following a period-doubling
bifurcation sequence.
|
